Developmental Psychology © 2010 American Psychological Association

2010, Vol. 46, No. 1, 46 –56 0012-1649/10/$12.00 DOI: 10.1037/a0017420

Adolescent Peer Relationships and Behavior Problems Predict Young

Adults’ Communication on Social Networking Websites

Amori Yee Mikami, David E. Szwedo, Joseph P. Allen, Meredyth A. Evans, and Amanda L. Hare
University of Virginia

This study examined online communication on social networking web pages in a longitudinal sample of 92
youths (39 male, 53 female). Participants’ social and behavioral adjustment was assessed when they were ages
13–14 years and again at ages 20 –22 years. At ages 20 –22 years, participants’ social networking website use
and indicators of friendship quality on their web pages were coded by observers. Results suggested that youths
who had been better adjusted at ages 13–14 years were more likely to be using social networking web pages
at ages 20 –22 years, after statistically controlling for age, gender, ethnicity, and parental income. Overall,
youths’ patterns of peer relationships, friendship quality, and behavioral adjustment at ages 13–14 years and
at ages 20 –22 years predicted similar qualities of interaction and problem behavior on their social networking
websites at ages 20 –22 years. Findings are consistent with developmental theory asserting that youths display
cross-situational continuity in their social behaviors and suggest that the conceptualization of continuity may
be extended into the online domain.

Keywords: online, social networking, adolescents, friendship, peer relationships

Explosive growth has occurred in online social communication Developmental Context of Social Communication
(Madden, 2006), with youths disproportionately affected by this
Characterizations of youths’ social relationships in the internet
new technology (Pew Internet and American Life Project, 2009).
medium, as well as the investigation of continuity between face-
As online use increases, so too do debates about how internet-
to-face and online social behaviors, carry high relevance for de-
based interaction may compare with historical face-to-face ways of
velopmental psychology. It is during the adolescent period that
communicating (Bargh & McKenna, 2004; Tyler, 2002). One
peer interactions arguably hold the greatest importance for indi-
argument posits that internet interaction is often of lower quality
viduals’ social and behavioral functioning (Berscheid, 2003; Col-
than is face-to-face interaction, because constraints inherent in the
lins, 1997; Gifford-Smith & Brownell, 2003). The quantity of peer
online medium hinder relationships. Furthermore, use of online
interactions and the intimacy in friendships rise dramatically dur-
communication may be positively correlated with adjustment
ing this time (Berndt, 1999; Furman & Buhrmester, 1992); this
problems because (a) socially inept youths are drawn to online
increase in emotional closeness may be especially relevant for the
interaction and (b) the almost inevitably poor quality of online
friendships of girls (Maccoby, 1998). The correlation between
communication increases maladjustment. An alternative argument
friendship quality and adjustment (Buhrmester, 1990), as well as
postulates that the internet is merely a new medium for youths to
the influence of peers’ behavior on youths’ own behavior through
display the same long-standing patterns as they do using modes
contagion effects (Dishion & Owen, 2002; Harris, 1995), is also
other than online forms of communication, such that there is
suggested to peak during adolescence. Increasingly, a significant
correspondence between face-to-face and online interaction styles
proportion of these important peer interactions occurs online for
and friendship quality. In contrast to the first argument, use of
many adolescents (Bargh & McKenna, 2004), making them rele-
online communication may be negatively correlated with adjust-
vant for researchers to characterize.
ment problems, because socially competent youths treat the online
The templates that peer relationships establish in early adoles-
environment as yet another place in which to interact with existing
cence may further become critical in early adulthood as peers
friends and broaden their social circle.
become primary sources of support. Existing research suggests
continuity in patterns of interpersonal communication and relation-
Amori Yee Mikami, David E. Szwedo, Joseph P. Allen, Meredyth A. ship quality over time and across social contexts. Several long-
Evans, and Amanda L. Hare, Department of Psychology, University of term, longitudinal studies find that characteristics of peer interac-
Virginia. tions in childhood and early adolescence are repeated in young
This study was supported by National Institute of Mental Health Grants adult relationships with romantic partners (Sroufe, Egeland, Carl-
1R01MH44934 and 1R01MH58066. We thank the participants and their son, & Collins, 2005; Stocker & Richmond, 2007) and friends
friends and families. We are grateful to the many individuals who helped
(Bagwell, Newcomb, & Bukowski, 1998; Eisenberg et al., 2002).
collect data for this study, in particular Anne Dawson, Joanna Chango,
Megan Schad, April Reeves, Claire Stephenson, and Jessica VanAtta.
Similarly, externalizing and internalizing behavior problems dis-
Correspondence concerning this article should be addressed to Amori played in social relationships at one time point have been shown to
Yee Mikami, 102 Gilmer Hall Box 400400, Department of Psychology, repeat themselves in relationships several years later (e.g., Ped-
University of Virginia, Charlottesville, VA 22904-4400. E-mail: ersen, Vitaro, Barker, & Borge, 2007; Stocker & Richmond,
Mikami@virginia.edu 2007). Given this evidence for continuity in social interaction,

46
 ONLINE SOCIAL NETWORKING
early adolescents’ patterns of face-to-face communication with
peers may be replicated in an online medium as young adults.
Quality of Internet-Based Relative to Face-to-Face
Interaction
Existing research about online social communication remains
limited. To our knowledge, all published, empirical studies to date
have correlated individuals’ self-reports of their online relation-
ships and adjustment at a single time point, with the exception of
Kraut et al. (1998, 2002), who used objective measures of internet
use and self-report measures of adjustment in a 3-year longitudinal
design, and Willoughby (2008), who used self-report measures in
a 21-month design. Whether earlier patterns of peer interactions
show continuity with behavior in the online medium remains
largely unknown. Further, observational data about friendship
quality online is completely lacking. Herein, we review support
for the first argument, that internet-based interaction is of poorer
quality than face-to-face interaction, and then review support for
the second argument, that the internet represents a new medium for
youths to enact their long-standing patterns of social communica-
tion displayed face-to-face.
Theoretical reasons have been proposed stating that because
nonverbal cues and personalizing information are limited online,
internet-based interactions result in lower quality relationships
than do face-to-face interactions (Keisler, Seigel, & McGuire,
1984). Two studies that compared participants’ self-report of their
internet and face-to-face relationships found that youths rarely
become as close to online friends as they do to in-person friends
(Mesch & Talmud, 2007; Parks & Roberts, 1998). Further argu-
ments have been proposed that internet use contributes to poor
adjustment because youths’ online social interactions do not sub-
stitute for (or potentially take time away from) intimate face-to-
face relationships. Some research has found youths’ self-report of
internet use to be cross-sectionally correlated with self-report of
adjustment problems (Modayil, Thompson, & Varnhagen, 2003;
Morahan-Martin & Schumacher, 2003; Ybarra, Alexander, &
Mitchell, 2005) and poor face-to-face relationships (Papacharissi
& Rubin, 2000; Sanders, Field, Diego, & Kaplan, 2000). However,
the direction of effects in these studies is unknown. It may be that
poorly adjusted youths are drawn to internet communication be-
cause of their failure in face-to-face relationships (Wolak, Mitch-
ell, & Finkelhor, 2003) and that such youths would have difficulty
establishing positive social ties in any venue. However, consistent
with the hypothesized causal pathway, Kraut et al. (1998, 2002)
found that adolescents’ objectively assessed internet use predicted
self-reports of decreased social support and increased psychopa-
thology over a 1- to 2-year period, although these effects disap-
peared at a subsequent follow-up.
In support of the alternative viewpoint, some research indicates
that the characteristics of individuals’ face-to-face relationships
play out again in the online medium. Adolescents who described
themselves as having internalizing symptoms (Gross, Juvonen, &
Gable, 2002; Ybarra et al., 2005) were more likely to communicate
over the internet with people they knew less well, and to talk about
superficial topics, than were adolescents without internalizing
symptoms. In retrospective self-report designs, individuals’ prob-
lematic internet interactions have appeared to be extensions of
difficulties that predated their internet use (Mitchell, Finkelhor, &
47
Becker-Blease, 2007; Modayil et al., 2003). Similarly, youths with
strong, positive face-to-face relationships may be those most fre-
quently using internet social communication as an additional
venue in which to interact. In support of this argument, work has
found youths’ self-report of internet use to be cross-sectionally
correlated with self-reports of extraversion (Peter, Valkenburg, &
Schouten, 2005), low social anxiety (Valkenburg & Peter, 2007),
and sociability in face-to-face interactions (Birnie & Horvath,
2002). Willoughby (2008) found trends that adolescents’ self-
reports of better relationships with peers, though not with parents,
predicted increases in internet use over a 21-month period.
Social networking websites, the most popular of which are
Facebook and MySpace, exemplify online social communication.
Because of their recency, few studies have specifically investi-
gated these explicitly socially oriented websites as opposed to the
research reviewed studying other types of internet use. Notably,
unlike other forms of internet communication, these sites encour-
age nonanonymous interactions and the recognition of connections
between friends. Ellison, Steinfield, and Lampe (2007) found that
the amount of time college students reported using Facebook was
positively correlated with their self-reported face-to-face involve-
ment in the college community; this relationship held after statis-
tical control of total internet use, suggesting a unique function of
Facebook to enhance social communication. In a study of a Dutch
social networking website, Valkenburg, Peter, and Schouten
(2006) found that adolescents who self-reported receiving positive
comments from friends posted on their page also self-reported
good adjustment. Walther, Van Der Heide, Kim, Westerman, and
Tong (2008) found that participants judged Facebook page owners
on the basis of characteristics of the friends on the owners’ pages,
suggesting that youths view Facebook interactions as reflecting the
quality of owners’ face-to-face relationships.
Demographic Moderators of Online Social
Communication
Although historically adolescent boys have adopted internet
technology faster than girls, current research suggests gender
equivalence in most types of internet use (Gross, 2004; Joiner et
al., 2005; Subrahmanyam, Greenfield, Kraut, & Gross, 2001;
Willoughby, 2008) and potentially female predominance in the
social communication functions of the internet (Jackson, Ervin,
Gardner, & Schmitt, 2001a). Females may make more emotionally
supportive comments online than do males (Baron, 2004; Hartsell,
2005), which has been postulated to reflect gender differences also
displayed in face-to-face relationships. White and high-income
individuals may be more likely to use the internet than their
low-income or minority peers (Jackson, Ervin, Gardner, &
Schmitt, 2001b; Pew Internet and American Life Project, 2008),
but it is unknown how income or ethnicity might affect commu-
nication or relationship quality online.
Study Aims
The current study utilized observational measures and multiple
informants in a longitudinal design to examine predictors of young
adults’ communication on the social networking websites Face-
book and MySpace. We first hypothesized that youths with social
networking websites at ages 20 –22 years relative to youths with-
48 MIKAMI, SZWEDO, ALLEN, EVANS, AND HARE
out such websites would be those who were best adjusted (a) at
ages 13–14 years and (b) concurrently at ages 20 –22 years.
Among youths with web pages, we next hypothesized that youths’
adjustment, both at ages 13–14 years and at ages 20 –22 years,
would predict similar patterns of communication on their web
pages. Specifically, popularity and friendship quality at 13–14 and
20 –22 years of age were expected to predict social competence on
web pages, whereas problem behaviors at ages 13–14 and 20 –22
years would also be displayed online. We made this prediction on
the basis of developmental theory showing consistency in social
behaviors across medium and time. On the basis of established
gender differences in communication (Baron, 2004; Hartsell,
2005), we further hypothesized that girls would show more emo-
tionally supportive interactions on web pages than would boys. In
sum, we expected that data would support the theory that the
internet represents a new medium to display the same patterns of
social interaction as have been displayed in other venues.
Method
Participants were 92 (39 male, 53 female) youths taking part in
an ongoing longitudinal study of adolescent and young adult
development. The youths entered the study in 1998 –1999 when
they were in the seventh and eighth grades (mean age ⫽ 13.30
years, SD ⫽ 0.62) and were reassessed in 2006 –2008 as young
adults (mean age ⫽ 20.92 years; SD ⫽ 1.11). These youths are the
individuals for whom their social networking web pages could be
coded, and they represent a subset of the full sample of 184
participants assessed at baseline and 172 participants assessed at
follow-up.
At baseline, adolescents were recruited from a single public
middle school; the students had attended the same school as an
intact group since the fifth grade. Racial/ethnic diversity was as
follows: 58% White, 29% African American, and 13% other or
mixed. Annual family income was distributed as follows: 18%
reported incomes under $19,999; 28% had incomes of $20,000 –
$39,999; 22% reported incomes of $40,000 –$59,000; and 33%
had an income of $60,000 or more. Active, written informed
consent was obtained from participants at every assessment point;
if participants were minors, they provided assent, and their parents
provided consent. Study procedures were approved by a university
review board. For further details about the sample, please see
Allen, Porter, McFarland, Marsh, and McElhaney (2005).
Baseline Measures (Ages 13–14 Years)
Using a multimethod procedure, we assessed adolescents’ peer
status, friendship quality, and problem behaviors. These particular
peer relationship measures were selected because sociometric sta-
tus and friendship quality are believed to be distinct, yet important
ways of characterizing youths’ social competence (Parker &
Asher, 1993). Similarly, the adjustment measures chosen were
thought to reflect the central areas of problem behavior— both
internalizing and externalizing domains— known to relate to social
functioning (Asarnow, 1988). We note that baseline data collection
predated the advent of social networking websites, and therefore
social networking information was not available at this time point.
Peer sociometric status. We assessed peer sociometric status
using a limited nomination sociometric procedure adapted from
Coie, Dodge, and Coppotelli (1982) and modified for adolescents
(Franzoi, Davis, & Vasquez-Suson, 1994). Each participant nom-
inated up to 10 peers in their grade with whom they would most
like to spend free time outside of school and 10 peers with whom
they would least like to spend such free time. Because the entire
sample attended the same school, each adolescent’s nominations
came from 72 to 146 peers, depending on the adolescent’s grade
cohort. Sociometric status was calculated for each participant by
taking the number of “most liked” nominations received minus the
number of “least liked” nominations received, divided by the
number of peers making nominations. This procedure has yielded
stability over a 1-year period (r ⫽ .78 for positive and r ⫽ .66 for
negative nominations), strong links to friendship and attachment se-
curity, and longitudinal links to decreasing levels of hostility over
time (Allen et al., 2005; Allen, Porter, McFarland, McElhaney, &
Marsh, 2007).
Observed positivity and negativity in peer interaction.
Each adolescent and his or her best friend (nominated by the
adolescent) participated in a dyadic interaction in which the pair
decided together who, out of a hypothetical cast of characters,
should be selected to be rescued from a desert island. The task was
intended to produce discussion and conflict. Coders, unaware of
other data about the participants, rated the positivity (smiling
genuinely toward one another, validating the other person’s ideas,
making truly friendly jokes, and listening to the other person) and
negativity (interrupting the other person, criticizing the other per-
son’s ideas, and appearing uninterested in what the other person
was saying) in the dyadic interaction. Positive and negative indi-
cators were considered separately as has been recommended by
other researchers (Hartup, 1995). This standardized coding system
has demonstrated validity (Allen, Porter, & McFarland, 2006;
Allen et al., 2007). Interrater reliability between raters was accept-
able (intraclass correlation coefficients [ICCs]: .65–.86).
Depressive symptoms. Adolescents completed the Child De-
pression Inventory (CDI; Kovacs & Beck, 1977), a widely used
self-report measure of depressive symptoms (Twenge & Nolen-
Hoeksema, 2002). We chose a self-report measure because of
consensus that adolescents are the most valid reporters of their
own internalizing problems (Bird, Gould, & Staghezza, 1992). The
CDI has been documented to relate to clinical depression among
adolescents (Craighead, Curry, & Ilardi, 1995; Timbremont &
Braet, 2004). The scale contains 27 items (sample item: “I am sad
once in a while”), each rated on a metric from 0 to 2 to indicate the
youth’s feelings within the past 2 weeks. Alpha level in our sample
was .86.
Delinquent behavior. Primary caregivers (98% of whom
were mothers) completed the short form of the Delinquency scale
of the Child Behavior Checklist (Achenbach, 1991), which as-
sesses the adolescent’s delinquent and acting out behavior (sample
items: “She destroys things belonging to her family or others”;
“She is disobedient at school”). Each item is rated on a metric from
0 to 2 and summed to produce a composite score. Reliability and
validity of this form in relation to the full scale has been docu-
mented (Lizotte, Chard-Wierschem, Loeber, & Stern, 1992). All
adolescents were living with the primary caregiver who completed
the report. We relied on parental report to assess these behaviors
because of evidence that adults are the most valid informants of
disruptive behavior (Loeber, Green, Lahey, & Stouthamer-Loeber,
1991). Alpha in our sample for the six-item scale was .69.
 ONLINE SOCIAL NETWORKING
Online Social Communication Follow-up Measures
When participants were young adults, we assessed online social
networking web page use and friendship quality on these web
pages, as well as indicators of adjustment. Of the 172 participants
reached at the follow-up assessment point, 147 reported that they
had a social networking web page; of these, 118 gave consent to
code their page, and we successfully coded 92 of these pages. The
pages we could not code (a) were on a social networking website,
such as Friendster, with a scheme that was too different from
Facebook or MySpace for us to apply the coding; (b) could not be
located; or (c) were not accessible because the participant con-
sented during the visit but failed to accept our “friend request” on
the website (which is needed to view the page). There were no
significant differences between youths who gave consent for us to
code their web pages versus youths with web pages who did not
give consent on any baseline measure: age, t(145) ⫽ ⫺0.10;
family income, t(145) ⫽ ⫺0.75; gender, ␹2(1, N ⫽ 147) ⫽ 0.07;
ethnic minority status, ␹2(1, N ⫽ 147) ⫽ 1.21; peer sociometric
status, t(145) ⫽ ⫺0.85; observed positivity, t(145) ⫽ 0.47, or
negativity, t(145) ⫽ ⫺0.81, in peer interaction; delinquent behav-
ior, t(145) ⫽ 0.36; or depressive symptoms, t(145) ⫽ ⫺1.57, all ps
⬎.05. Additionally, there were no differences between the youths
whose pages we successfully coded and the youths with web pages
that we did not code.
Facebook and MySpace are by far the most popular social
networking websites (TechCrunch, 2008), which is why we chose
them for coding. Of the pages coded in our sample, 66% were on
Facebook and 34% were on MySpace. On these websites, users
create their own page where they typically include their name,
photos, and information about themselves. Individuals are linked
to friends in a social network, and friends post comments on each
others’ pages that may be viewed by all network members. For
coding, we selected indicators reflecting the most prominent fea-
tures of these web pages: the description of the page owner
(hostility in “About Me” sections, inappropriateness in photos
posted by the owner) and the characteristics of the friends (number
of friends and connection and support in friends’ posts on the
page). Seventy pages were selected at random to be double coded
to provide an estimate of consistency among raters.
Number of friends on web page. Coders recorded the total
number of friends in the participant’s social network, which is
listed on the web page (ICC ⫽ .97).
Connection with friends on web page. Coders examined the
most recent 20 posts from friends displayed on the participant’s
web page. This number was chosen because most participants had
many more than 20 posts, and the most recent 20 are automatically
displayed on a participant’s web page when someone clicks on the
link to view posts. Coders counted the number of different friends
who had posted comments among these 20 that indicated that the
friend and the participant shared an actual relationship. This con-
struct captured whether the participant was primarily communi-
cating online with existing friends or communicating with strang-
ers or distant acquaintances, as has been found in previous research
to characterize youths with psychopathology (Wolak et al., 2003).
Posts in which the friend said “see you at dinner” or “wasn’t that
lecture today hilarious?” suggested a connected relationship be-
tween individuals who communicate with one another outside of
the online realm. Posts that appeared to be chain letters or posts,
49
such as “what’s up?” were not counted as connection, as they were
not considered to demonstrate a close relationship between the
friend and the participant (ICC ⫽ .84). Friends’ connection posts
were positively correlated with participants’ self-report on the
Online Friendships Questionnaire that they “talk with people on-
line often seen in person” (r ⫽ .29; p ⫽ .03) and were negatively
correlated with self-report that they “have a close friendship with
someone they met online” (r ⫽ ⫺.35; p ⬍ .01), questions used and
validated in previous research (Morahan-Martin & Schumacher,
2003).
Support of friends on web page. Again in the most recent 20
posts from friends on the web page, coders recorded the number of
different friends making posts characterized by strong words of en-
couragement, compliments, understanding, caring, or validation. This
construct indicated supportive relationship quality on the website,
similar to ways in which emotionally supportive comments are coded
in face-to-face interaction tasks. Posts in which the friend said “I miss
you so much,” “I love you,” or “you’re truly my inspiration” would
exemplify support. Posts that were mildly positive in tone, such as
“take care” or “have a good weekend,” were not considered support
because these comments seem plausible in a casual relationship.
Comments without a validation, such as “I’m so happy to be going
home this weekend” or “did you do the lab homework?” were not
considered support, although the question about homework would be
considered connection (ICC ⫽ .83).
Hostility in “About Me” section. Participants’ descriptions
of themselves on the website were coded for the presence of
aggression, hostility, or negativity on a dichotomous metric. Cod-
ers assigned a 1 if there was no indication of these themes. A 2 was
coded if there was any mention of anger or hostility toward the
world, criticism or annoyance expressed, or explicit mention of
any threat of violence. For example, the statement “I live in the
worst town you could ever imagine, not only is it boring here, but
it is consumed by pricks” would be counted for the presence of
hostility (␬ ⫽ .63).
Inappropriate photos posted. After looking at all photos on
the website posted by the participant, coders assigned a rating of
1–3 to indicate the degree to which any picture showed behavior
that might be considered inappropriate if viewed by an authority
figure. To establish this construct, we researched what type of
photos posted on social networking websites lead employers to not
hire applicants (Finder, 2006). All participants had at least one
photo on their web page. Coders assigned a 1 if there were no
inappropriate photos displayed. Coders assigned a 2 for the pres-
ence of minor inappropriate behavior and a 3 for the presence of
significant inappropriate behavior in any photo. Benign photos of
the participant, including most photos of the participant drinking
beer, would be scored 1. A score of 2 was given for any picture of
downing shots of alcohol or the participant in provocative clothing.
If there was any photo of the participant naked, making obscene
gestures, or engaging in vandalism, the construct was scored 3
(ICC ⫽ .76).
Social and Behavioral Adjustment Follow-up Measures
We selected measures of social and behavioral adjustment in
face-to-face relationships at follow-up that were conceptually the
most similar to those included at baseline.
50 MIKAMI, SZWEDO, ALLEN, EVANS, AND HARE
Friendship quality. The age of the participants at the
follow-up assessment point precluded assessment of peer socio-
metric status, and observational measures of friendship quality
were not available. However, participants and their best friends
(nominated by the participant) each independently reported on the
quality of their friendship using the Network of Relationships
Inventory (NRI; Furman & Buhrmester, 1985), a 45-item scale
commonly used in this age group. We considered positive features
(e.g., companionship, intimacy, nurturance, affection) and nega-
tive features (e.g., conflict, antagonism, criticism, punishment) of
friendship quality separately. Reliability and validity have been
documented (Furman, 1996).
Depressive symptoms. Participants self-reported their de-
pressive symptoms on the Beck Depression Inventory–II (Beck,
Steer, & Brown, 1996), a widely used, 21-item scale with good
psychometric properties (Steer, Ball, Ranieri, & Beck, 1999) in
which participants endorse depressive symptoms experienced in
the past 2 weeks. Alpha in our sample was .85.
Rule-breaking behavior. Parental report was no longer ap-
propriate for this age group. We collected best friend reports of the
youths’ behavior on the Young Adult Behavior Checklist (Achen-
bach, 1997) Rule-Breaking subscale. This subscale contains 14
items. Alpha in our sample was .86.
Data Analytic Plan
We used binary logistic regression to test our first hypothesis
that positive adjustment (a) in early adolescence and (b) concur-
rently in young adulthood would predict having a social network-
ing web page in young adulthood. The presence versus absence of
a social networking web page was the criterion variable, and the
measures of friendship quality and adjustment at each age period
were the predictors. We included the covariates of participant
gender, age, and ethnic minority status as well as baseline family
income. Gender and minority status were dichotomous, dummy-
coded variables (1 ⫽ male, 2 ⫽ female; 1 ⫽ White, 2 ⫽ minority).
We next tested the hypotheses that among the group of youths
with social networking web pages, adjustment at 13–14 and 20 –22
years of age would be reflected in the youths’ web pages at 20 –22
years of age and that females would display more supportive
relationships on their web pages than would males. First, regarding
predictive relationships between early adolescent adjustment and
young adult online relationships, we conducted ordinary least
squares hierarchical multiple regressions for the continuous vari-
ables of (a) number of friends in the network, (b) friends posting
connection comments, and (c) friends posting support comments;
binary logistic regression for (d) hostility in the “About Me”
section; and ordinal logistic regression for (e) inappropriate pho-
tos. At Step 1, we entered the demographic covariates of baseline
family income, gender, age, and ethnic minority status. At Step 2,
we entered the measures of baseline social adjustment together: (a)
peer sociometric status, (b) observed positivity in peer interaction,
(c) observed negativity in peer interaction, (d) depressive symp-
toms, and (e) delinquent behavior. If the baseline measures of
sociometric status, positivity, and negativity in peer interaction
predicted the number of friends in network, friends posting con-
nection comments, and friends posting support comments, then the
hypothesis regarding continuity of social status and friendship
quality into the online domain would be confirmed. If the baseline
measures of depressive symptoms and delinquent behavior pre-
dicted hostility in the “About Me” section and inappropriate pho-
tos, then the hypothesis regarding continuity of behavior problems
into the online domain would be confirmed.
Because participants who had social networking web pages did
not randomly represent the full sample of youths, we used the
Heckman two-stage procedure to correct for selection bias (Heck-
man, 1979). In the Heckman correction, in the first stage, we used
probit regression to estimate the likelihood of having a web page
and, in the second stage, incorporated parameter estimates from
the selection equation to predict the relationship between early
adolescent adjustment and online communication. This procedure
reduced the likelihood of a bias in regression estimates resulting
from unmeasured differences between youths who have web pages
and youths who do not.
We entered all possible interactions between gender and base-
line social adjustment at Step 3 in each regression. Out of 20
possible interactions, two were significant. Probing revealed op-
posite effects such that the relationship between one measure of
adjustment and online communication was stronger for girls, but
the relationship between the other measure and online communi-
cation was stronger for boys. Because the number of significant
interactions was close to what would be expected by chance, and
because the directions of the two effects were inconsistent, we do
not report these findings in the current article.
Next, regarding correspondence between concurrent indicators
of adjustment and patterns of online social communication at
20 –22 years of age, we conducted correlations between these
measures. If social adjustment measures were correlated with a
greater number of friends, friends’ connection, and friends’ sup-
port online, and behavior problem measures were correlated with
hostility and inappropriate photos online, then this hypothesis
would be confirmed.
Results
Preliminary Analyses
All continuous study variables were roughly normally distrib-
uted, with the exception of number of friends on the web page,
which was positively skewed. We calculated the square root of this
variable, which yielded a normal distribution; we used the trans-
formed variable in analyses. We next examined the z-score distri-
butions of all variables. For no variable did any participant have a
z score greater than ⫾3.0. Descriptive statistics and correlations
among the baseline indicators of adjustment and follow-up social
networking variables are presented in Table 1. Variables within
each time period were modestly or moderately (and not highly)
correlated with one another, justifying the inclusion of each of
these separate indicators.
Factors Differentiating Youths With and Without
Social Networking Web Pages
Among the youths who completed follow-up measures (n ⫽
172), 147 (86%) reported that they had a personal social network-
ing web page, and 25 (14%) reported that they did not. We used
the full sample of participants with web pages for these analyses
(as opposed to the proportion of individuals with web pages who
 ONLINE SOCIAL NETWORKING 51

Table 1
Correlations Between Adjustment at 13–14 Years of Age and Online Communication at 20 –22 Years of Age

Variable Descriptive statistics 1 2 3 4 5 6 7 8 9 10

1. Sociometric status —
M 0.35
(SD) (1.65)
2. Observed positivity .12 —
M 2.36
(SD) (0.63)
3. Observed negativity ⫺.06 ⫺.05 —
M 0.78
(SD) (0.52)
4. Depressive symptoms ⫺.08 ⫺.07 .04 —
M 5.07
(SD) (4.28)
5. Delinquent behavior ⫺.27ⴱⴱ ⫺.18ⴱ .09 .18ⴱ —
M 1.52
(SD) (1.82)
6. Number of friends .25ⴱ .29ⴱⴱ ⫺.18 ⫺.11 ⫺.30ⴱⴱ —
M 298.60
(SD) (247.6)
7. Friends’ connection .27ⴱⴱ .41ⴱⴱ .10 ⫺.14 ⫺.37ⴱⴱ .42ⴱⴱ —
M 5.74
(SD) (2.64)
8. Friends’ support .28ⴱⴱ .10 ⫺.15 ⫺.14 ⫺.21ⴱ .29ⴱⴱ .24ⴱ —
M 1.63
(SD) (1.35)
9. Hostility in about me 1 ⫽ 79, 2 ⫽ 13 .00 ⫺.27ⴱⴱ .04 ⫺.16 .21ⴱ .04 ⫺.04 ⫺.01 —
(86%) (14%)
10. Inappropriate photos 1 ⫽ 66, 2 ⫽ 16, 3 ⫽ 10 .04 .08 ⫺.08 .21ⴱ .01 .02 .13 .01 ⫺.12 —
(72%) (17%) (11%)

Note. Means (with standard deviations in parentheses) are presented for raw scores on continuous variables. Pearson product moment correlations are
presented for continuous variables, point biserial correlations were presented for continuous with categorical variables, and phi coefficients were calculated
for categorical variables. Constructs 1–5 were assessed at baseline (13–14 years of age); N ⫽ 172. Constructs 6 –10 were assessed at follow-up (20 –22 years
of age); N ⫽ 92.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01.

gave permission for us to code them) as a more conservative test
of the hypothesis that better adjusted youths have web pages.
Having a social networking web page was not significantly
associated with the demographic factors of participant age, odds
ratio (OR) ⫽ 0.64, 95% confidence interval (CI) ⫽ 0.41–1.01, p ⫽
.06; ethnic minority status, OR ⫽ 0.72, CI ⫽ 0.25–2.11, p ⬎ .05;
gender, OR ⫽ 1.26, CI ⫽ 0.51–3.14, p ⬎ .05; or baseline family
income, OR ⫽ 1.51, CI ⫽ 0.93–2.45, p ⬎ .05. However, after
statistical control of demographic factors, early adolescent youths
who had displayed greater observed negativity in their face-to-face
peer interactions and more depressive symptoms were signifi-
cantly less likely to have a social networking webpage in young
adulthood. With regard to concurrent adjustment, young adults
who self-reported more positive features in their close friendships
were also more likely to have a social networking webpage (see
Table 2).
Longitudinal Associations Between Early Adolescent
Adjustment and Web Page Indicators
We conducted analyses limited to the subsample of youths with
social networking web pages that we coded (n ⫽ 92).
Number of friends on web page. At Step 1, we found that
having more friends was correlated with being a member of an
ethnic minority group. However there were no main effects for
gender, age, or baseline parental income. At Step 2, we found
significant effects for two of the baseline social adjustment mea-
sures in predicting the number of friends the participant listed on
their web page at 20 –22 years of age. Youths with higher early
adolescent positivity in their peer dyadic interaction had more
friends in their online network. Youths with more negativity in
their peer dyadic interaction, however, had fewer friends in their
online networks (see Table 3).
Connection of friends on web page. As displayed in Table 3,
at Step 1 we controlled for demographic variables, none of which
was associated with the criterion variable of number of friends
posting comments on the web page indicating connection (out of
the most recent 20 posts). At Step 2, observed positivity in peer
interactions predicted a larger number of individuals posting con-
nection comments. Parent-reported delinquency, however, pre-
dicted a smaller number of individuals posting connection com-
ments.
High support of friends on web page. As displayed in Table
3, at Step 1 participant gender was significant, such that female
participants, in comparison with male participants, had more
friends posting highly supportive comments on their web page.
However no other demographic variable predicted support. After
52 MIKAMI, SZWEDO, ALLEN, EVANS, AND HARE

Table 2
Early Adolescent Adjustment of Youths With Social Networking Web Pages in Young Adulthood

Has web page (n ⫽ 147) No web page (n ⫽ 25)

at 20–22 years of age at 20–22 years of age
95% Confidence
Variable M (SD) M (SD) Odds ratio interval

Ages 13–14 years

1. Sociometric status 0.47 (0.96) 0.07 (1.56) 1.31 0.78–2.20
2. Observed positivity 2.38 (0.63) 2.22 (0.51) 1.12 0.71–1.79
3. Observed negativity 0.74 (0.47) 0.99 (0.67) 0.68ⴱ 0.45–0.99
4. Depressive symptoms 4.68 (3.84) 7.44 (6.28) 0.63ⴱ 0.43–0.93
5. Delinquent behavior 1.35 (1.81) 2.38 (1.70) 0.72 0.47–1.10
Ages 20–22 years
1. Self-report positivity 11.88 (2.05) 10.24 (2.41) 2.05ⴱⴱ 1.26–3.35
2. Self-report negativity 4.31 (1.64) 4.47 (2.03) 1.10 0.70–1.71
3. Peer-report positivity 11.42 (1.91) 11.03 (2.40) 1.17 0.74–1.86
4. Peer-report negativity 4.46 (1.71) 4.34 (1.32) 1.12 0.64–1.96
5. Depressive symptoms 5.32 (6.08) 6.13 (7.34) 0.90 0.57–1.43
6. Rule breaking 3.68 (4.26) 3.93 (4.04) 0.95 0.58–1.54

Note. Presented means are raw scores (unadjusted for covariates), with standard deviations in parentheses. Odds ratios, however, were calculated by
logistic regression with statistical control of participant age at baseline, gender, ethnic minority status, and baseline family income.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01.

statistical control of income, gender, age, and ethnic minority
status, at Step 2 having higher sociometric status at 13–14 years of
age predicted more friends posting supportive comments on the
web page.
Hostility in “About Me” section. No demographic variable at
Step 1 predicted the presence of hostility in the participant’s “About
Me” section. At Step 2, participants who had higher parent-rated
delinquent behavior at 13–14 years of age were more likely to display
hostility in their “About Me” section (see Table 4).
Inappropriate photos posted. No demographic variable at
Step 1 was associated with the presence of inappropriate photos on
the web page. At Step 2, more self-reported depressive symptoms
at 13–14 years of age significantly predicted the presence of
inappropriate photos on web pages (see Table 4).
Concurrent Associations Between Young Adult
Adjustment and Online Social Communication
Table 5 presents correlations between concurrent indicators of
social adjustment and problem behavior, on the one hand, and
measures of online communication in young adulthood, on the
other. Youths’ self-report of positivity in their friendship was
positively correlated with number of friends and support of friends
on the web page. Youths’ self-report of negativity in their friend-

Table 3
Social Adjustment at 13–14 Years of Age Predicts Friendship on Social Networking Web Pages
at 20 –22 Years of Age

Predictor Number of Friends’ Friends’

(ages 13–14 years) friends ␤ connection ␤ support ␤

Step 1
Family income .23 .09 .14
Female gender ⫺.02 .09 .33ⴱⴱ
Ethnic minority .28ⴱ ⫺.14 ⫺.07
Age at baseline ⫺.12 ⫺.02 .02
Summary statistics
Total R2 .20ⴱ .20ⴱⴱ .24ⴱⴱ
⌬R2 .20ⴱ .20ⴱⴱ .24ⴱⴱ
Step 2
Sociometric status .18 .09 .24ⴱ
Observed positivity .32ⴱⴱ .29ⴱⴱ ⫺.04
Observed negativity ⫺.27ⴱ .20 ⫺.14
Delinquent behavior ⫺.13 ⫺.23ⴱ ⫺.03
Depressive symptoms ⫺.23 ⫺.02 ⫺.03
Summary statistics
Total R2 .32ⴱⴱ .35ⴱⴱ .29ⴱⴱ
⌬R2 .12ⴱ .16ⴱⴱ .05

Note. ⌬R2 effect size conventions: .01 ⫽ small; .06 ⫽ medium; .14 ⫽ large.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01.
 ONLINE SOCIAL NETWORKING
Table 4
Behavioral Adjustment at 13–14 Years of Age Predicts Problem Behaviors on Social Networking
Web Pages at 20 –22 Years of Age
Hostility
Predictor
(13–14 years of age) Odds ratio
Step 1
Family income 0.53
Female gender 0.77
Ethnic minority 0.56
Age at baseline 0.60
Step 2
Sociometric status 1.08
Observed positivity 0.48
Observed negativity 2.07
Delinquent behavior 2.73ⴱ
Depressive symptoms 1.16
ⴱ
p ⬍ .05.
ship was negatively correlated with connection of friends on the
web page. Best friend reports of positivity and negativity were
associated, positively and negatively, respectively, with observed
support on the web page. With regard to problem behaviors,
concurrent rule-breaking behavior, but not depression, was corre-
lated with inappropriate photos. Neither problem behavior was
significantly correlated with hostility on the web pages.
Discussion
Findings from this study suggest that youths may use social
networking websites to enact their long-standing face-to-face pat-
terns of interaction. After accounting for demographic factors, we
found that youths who were better adjusted at 13–14 years of age,
as indicated by having less observed negativity in face-to-face peer
interactions and fewer self-reported depressive symptoms, were
more likely to be using social networking websites at 20 –22 years
of age. In addition, youths at age 20 –22 who self-reported more
positivity in their closest friendship were more likely to be using
social networking websites.
Among the group of youths with social networking web
pages, the presence of higher positivity and lower negativity in
a dyadic peer interaction in early adolescence each predicted a
greater number of friends on the website. Higher observed
positivity in the peer interaction and lower parent-reported
delinquent behaviors predicted a greater number of friends
posting comments indicating connection. Higher peer sociomet-
ric status at baseline predicted a greater number of friends
posting comments indicating support. Early adolescent delin-
quent behaviors predicted the presence of hostility in the par-
ticipant’s “About Me” section, and depressive symptoms pre-
dicted the presence of photos on the website that might be
considered inappropriate. Females were more likely to have
friends posting supportive comments than were males, but the
predictive relationships between early adolescent adjustment
and young adult online communication did not differ by gender.
Analyses further suggested that concurrent social and behav-
ioral adjustment was also associated with similar patterns of
communication online. Self- and peer-reported positive friend-
53
Inappropriate photos
95% Confidence 95% Confidence
interval Odds ratio interval
0.22–1.29 0.79 0.51–1.23
0.19–3.08 0.80 0.25–2.56
0.29–1.10 0.80 0.15–4.29
0.27–1.36 0.80 0.43–1.49
0.40–2.94 0.98 0.65–1.49
0.21–1.10 1.05 0.44–2.54
0.49–8.71 1.02 0.20–5.10
1.05–7.13 0.97 0.65–1.44
0.24–5.66 1.27ⴱ 1.02–1.57
ship quality was concurrently associated with more friends
posting supportive comments online; self-reported positivity
was also associated with a larger number of friends. Self-
reported negativity in the friendship predicted fewer friends
posting comments indicating connection; peer-reported nega-
tivity predicted fewer friends posting supportive comments.
Concurrent peer reports of the participants’ rule-breaking be-
havior was associated with more posted pictures of the partic-
ipant engaging in inappropriate actions.
Taken together, these results are consistent with existing devel-
opmental theory that youths display cross-situational continuity in
their interpersonal interactions and suggest that the conceptualiza-
tion of continuity may be extended into the online domain. Find-
ings also highlight the potential importance of establishing positive
peer relationships in the developmental period between early ad-
olescence and young adulthood, suggesting that sociometric status
and friendship quality at 13–14 years of age may set the stage for
youths’ relationship quality at 20 –22 years of age. Further, indi-
cators of delinquent and depressive psychopathology in early
adolescence may manifest themselves in young adult relationships,
suggesting that adjustment patterns in early adolescence can carry
long-lasting effects.
Our findings diverge from previous studies suggesting low
correspondence between online and face-to-face relationships and
that maladjusted youths are drawn to the internet. However, pre-
vious studies were conducted when far fewer households had
access to online technology. Given the high degree to which social
networking web pages are used by youths today, it is more likely
that this represents a normative way of communicating. Social
networking web pages also differ from previous internet commu-
nication tools (such as instant messaging and e-mail) in that they
support communication with many friends and encourage users to
recognize connections between individuals. The graphical inter-
face of the web page allows users to share information with friends
in a vivid, easily accessible manner. In short, the web has changed
rapidly enough that studies based on usage from the year 2000 may
not reflect the impact of sites explicitly designed to facilitate social
networking on youths’ interactions.
54 MIKAMI, SZWEDO, ALLEN, EVANS, AND HARE

Table 5
Correlations Between Concurrent Adjustment and Online Communication at 20 –22 Years of Age

Variable Descriptive statistics 1 2 3 4 5 6 7 8 9 10 11

1. Youth self-report positivity —

M 11.61
(SD) (2.28)
2. Youth self-report negativity .06 —
M 4.36
(SD) (1.69)
3. Peer report positivity .31ⴱⴱ ⫺.02 —
M 11.39
(SD) (1.97)
4. Peer report negativity .13 .36ⴱⴱ ⫺.09 —
M 4.47
(SD) (1.69)
5. Youth self-report depressive
symptoms ⫺.04 .12 .00 ⫺.04 —
M 5.36
(SD) (6.22)
6. Peer report rule breaking ⫺.03 .02 ⫺.14 .46ⴱⴱ .07 —
M 3.72
(SD) (4.21)
7. Number of friends .28ⴱⴱ ⫺.07 .06 ⫺.02 ⫺.12 ⫺.05 —
M 298.60
(SD) (247.6)
8. Friends’ connection .15 ⫺.28ⴱⴱ .04 ⫺.06 ⫺.03 .02 .34ⴱⴱ —
M 5.74
(SD) (2.64)
9. Friends’ support .27ⴱⴱ ⫺.06 .28ⴱⴱ ⫺.27ⴱⴱ .12 ⫺.11 .27ⴱⴱ .24ⴱⴱ —
M 1.63
(SD) (1.35)
10. Hostility in about me 1 ⫽ 79, 2 ⫽ 13 .16 .17 .00 .14 .08 ⫺.05 .03 ⫺.04 ⫺.01 —
(86%) (14%)
11. Inappropriate photos 1 ⫽ 66, 2 ⫽ 16, 3 ⫽ 10 .02 ⫺.15 .16 .02 ⫺.11 .33ⴱⴱ .01 .14 .01 ⫺.10 —
(72%) (17%) (11%)

Note. Means (with standard deviations in parentheses) are presented for raw scores on continuous variables. Pearson correlations are presented for
continuous variables, point biserial correlations are presented for continuous with categorical variables, and phi coefficients are presented for categorical
variables. All constructs were assessed at follow-up (20 –22 years of age). N ⫽ 92.
ⴱⴱ
p ⬍ .01.

Nonetheless, our finding that the best adjusted youths (both in
early adolescence and in young adulthood) were those using online
social communication as young adults may specifically pertain to
social networking websites and not other online activities. Perhaps
youths with psychopathology were spending large amounts of time
online, but in other media not assessed as part of this study. We
speculate that well-adjusted youths may use social networking
websites in which visible communication is prominent, whereas
inept youths may prefer online activities where they have greater
anonymity. This hypothesis is supported by findings that people
with large face-to-face social networks are more likely to use the
internet to communicate with friends and family and less likely to
use the internet to communicate with strangers (Bessiere, Kiesler,
Kraut, & Boneva, 2008).
Strengths of this study include the longitudinal design and
observational measures of friendship quality, both in face-to-face
relationships and online. There is nearly complete independence of
method variance in primary analyses. The use of peer nominations
to assess sociometric status is rare, particularly in a study with
adolescent participants. Further, social networking websites repre-
sent a novel, observational medium for assessing peer relationships
that has high ecological validity.
A significant limitation of this study is the lack of online relation-
ship measures at the early adolescent assessment point. Because social
networking websites were started very recently (Facebook and
MySpace, e.g., were both founded in 2004 but did not achieve
widespread adoption until later), it was not possible to assess this type
of online social communication when the participants entered the
study in 1998 –1999. In addition, in 1998 –1999, it was uncommon for
youths in early adolescence, the age of the participants in this study,
to be using e-mail, instant messaging, or chat rooms; at that time,
these types of online communication tools were predominantly used
by college students. Nonetheless, our study was not able to control for
online communication use at baseline, which limits the conclusions
that may be drawn. In the future, we will be able to use indicators of
online social communication in the important objectives of prospec-
tively predicting changes in youths’ social and behavioral adjustment.
We speculate that online use in and of itself will not contribute to
adjustment but rather that positive social communication and friend-
ship quality online will further enhance social– emotional functioning,
whereas negative online relationships will predict increasing adjust-
ment problems.
Another limitation of the present study was the relatively small
sample size for the analyses limited to the participants with coded
 ONLINE SOCIAL NETWORKING
web pages. The high percentage of participants who did not give
permission to code their page raises the concern that the sample
may not reflect the population of social networking web page
users, although we did not find any evidence that individuals who
consented differed from those who did not consent on any of the
nine baseline measures of demographics and adjustment. A further
limitation is that the baseline sociometric measure, modified for
use with an adolescent sample, may not fully capture the peer
acceptance of youths in the way that such measures do for ele-
mentary school–aged children.
In sum, although earlier studies concluded that online social
relationships were nearly always of poorer quality than face-to-
face relationships, the results from the current study instead sug-
gest continuity between youths’ face-to-face and online commu-
nication patterns, friendship quality, and behavioral adjustment. In
a well-known cartoon for The New Yorker (Steiner, 1993), a dog in
front of a computer says to his canine companion, “on the Internet,
nobody knows you’re a dog” (p. 61). On the basis of the current
findings, however, it is perhaps more accurate to say “on the
internet, you behave like the dog that you are.”
References
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist and
Revised Child Behavior Profile. Burlington, VT: University of Vermont,
Department of Psychiatry.
Achenbach, T. M. (1997). Manual for the Young Adult Self-Report and
Young Adult Behavior Checklist. Burlington, VT: University of Ver-
mont, Department of Psychiatry.
Allen, J. P., Porter, M. R., & McFarland, F. C. (2006). Leaders and
followers in adolescent close friendships: Susceptibility to peer influ-
ence as a predictor of risky behavior, friendship instability, and depres-
sion. Development and Psychopathology, 18, 155–172.
Allen, J. P., Porter, M. R., McFarland, F. C., Marsh, P., & McElhaney,
K. B. (2005). The two faces of adolescent’ success with peers: Adoles-
cent popularity, social adaptation, and deviant behavior. Child Develop-
ment, 76, 747–760.
Allen, J. P., Porter, M. R., McFarland, F. C., McElhaney, K. B., & Marsh,
P. (2007). The relation of attachment security to adolescents’ paternal
and peer relationships, depression, and externalizing behavior. Child
Development, 78, 1222–1239.
Asarnow, J. R. (1988). Peer status and social competence in child psychi-
atric inpatients: A comparison of children with depressive, externalizing,
and concurrent depressive and externalizing disorders. Journal of Ab-
normal Child Psychology, 16, 151–162.
Bagwell, C. L., Newcomb, A. F., & Bukowski, W. M. (1998). Preadoles-
cent friendship and peer rejection as predictors of adult adjustment.
Child Development, 69, 140 –153.
Bargh, J. A., & McKenna, K. Y. A. (2004). The internet and social life.
Annual Review of Psychology, 55, 573–590.
Baron, N. S. (2004). See you online: Gender issues in college student use
of instant messaging. Journal of Language and Social Psychology, 23,
397– 423.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck
Depression Inventory–II. San Antonio, TX: Psychological Corporation.
Berndt, T. J. (1999). Friends’ influence on students’ adjustment to school.
Educational Psychologist, 34, 15–28.
Berscheid, E. (2003). The human’s greatest strength: Other humans. In
L. G. Aspinwall & U. M. Staudinger (Eds.), A psychology of human
strengths: Fundamental questions and future directions for a positive
psychology (pp. 37– 47). Washington, DC: American Psychological As-
sociation.
55
Bessiere, K., Kiesler, S., Kraut, R., & Boneva, B. (2008). Effects of
internet use and social resources on changes in depression. Information,
Communication, and Society, 11, 47–70.
Bird, H. R., Gould, M. S., & Staghezza, B. M. (1992). Aggregating data
from multiple informants in child psychiatry epidemiological research.
Journal of the American Academy of Child and Adolescent Psychiatry,
31, 78 – 84.
Birnie, S. A., & Horvath, P. (2002). Psychological predictors of internet
social communication. Journal of Computer-Mediated Communication,
7(4).
Buhrmester, D. (1990). Intimacy of friendship, interpersonal competence,
and adjustment in preadolescence and adolescence. Child Development,
61, 1101–1111.
Coie, J. D., Dodge, K. A., & Coppotelli, H. (1982). Dimensions and types
of social status: A cross-age perspective. Developmental Psychology, 18,
557–570.
Collins, W. A. (1997). Relationships and development during adolescence:
Interpersonal adaptation to individual change. Personal Relationships, 4,
1–14.
Craighead, W. E., Curry, J. F., & Ilardi, S. S. (1995). Relationship of
Children’s Depression Inventory factors to major depression among
adolescents. Psychological Assessment, 7, 171–176.
Dishion, T. J., & Owen, L. D. (2002). A longitudinal analysis of friend-
ships and substance use: Bidirectional influence from adolescence to
adulthood. Developmental Psychology, 38, 480 – 491.
Eisenberg, N., Guthrie, I. K., Cumberland, A., Murphy, B. C., Shepard,
S. A., Zhou, Q., et al. (2002). Prosocial development in early adulthood:
A longitudinal study. Journal of Personality and Social Psychology, 82,
993–1006.
Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Face-
book “friends”: Social capital and college students’ use of online social
network sites. Journal of Computer-Mediated Communication, 12(4),
Article 1. Retrieved from http://jcmc.indiana.edu/vol12/issue4/
ellison.html
Finder, A. (2006, June 11). For some, online persona undermines a resume.
The New York Times. Retrieved from http://www.nytimes.com/2006/06/
11/us/11recruit.html?_r⫽1&pagewanted⫽all&oref⫽slogin
Franzoi, S. L., Davis, M. H., & Vasquez-Suson, K. A. (1994). Two social
worlds: Social correlates and stability of adolescent status groups. Jour-
nal of Personality and Social Psychology, 67, 462– 473.
Furman, W. (1996). The measurement of friendship perceptions: Concep-
tual and methodological issues. In W. M. Bukowski, A. F. Newcomb, &
W. W. Hartup (Eds.), The company they keep: Friendship in childhood
and adolescence. New York: Cambridge University Press.
Furman, W., & Buhrmester, D. (1985). Children’s perceptions of the
personal relationships in their social networks. Developmental Psychol-
ogy, 21, 1016 –1024.
Furman, W., & Buhrmester, D. (1992). Age and sex differences in per-
ceptions of networks of personal relationships. Child Development, 63,
103–115.
Gifford-Smith, M. E., & Brownell, C. A. (2003). Childhood peer relation-
ships: Social acceptance, friendships, and peer networks. Journal of
School Psychology, 41, 235–284.
Gross, E. F. (2004). Adolescent internet use: What we expect, what teens
report. Applied Developmental Psychology, 25, 633– 649.
Gross, E. F., Juvonen, J., & Gable, S. L. (2002). Internet use and well-being
in adolescence. Journal of Social Issues, 58, 75–90.
Harris, J. R. (1995). Where is the child’s environment? A group socializa-
tion theory of development. Psychological Review, 102, 458 – 489.
Hartsell, T. (2005). Who’s talking online: A descriptive analysis of gender
and online communication. International Journal of Information and
Communication Technology Education, 1, 42–54.
Hartup, W. W. (1995). The three faces of friendship. Journal of Social and
Personal Relationships, 12, 569 –574.
56 MIKAMI, SZWEDO, ALLEN, EVANS, AND HARE
Heckman, J. J. (1979). Sample selection bias as a specification error.
Econometrica, 47, 151–161.
Jackson, L. A., Ervin, K. S., Gardner, P. D., & Schmitt, N. (2001a). Gender
and the internet: Women communicating and men searching. Sex Roles,
44, 363–379.
Jackson, L. A., Ervin, K. S., Gardner, P. D., & Schmitt, N. (2001b). The
racial digital divide: Motivational, affective, and cognitive correlates of
internet use. Journal of Applied Social Psychology, 31, 2019 –2046.
Joiner, R., Gavin, J., Duffield, J., Brosnan, M., Crook, C., Durndell, A., et
al. (2005). Gender, internet identification, and internet anxiety: Corre-
lates of internet use. Cyberpsychology and Behavior, 8, 371–378.
Keisler, S., Seigel, J., & McGuire, T. W. (1984). Social and psychological
aspects of Computer Mediated Communication. American Psychologist,
39, 1123–1134.
Kovacs, M., & Beck, A. T. (1977). An empirical clinical approach toward
a definition of childhood depression. New York: Raven Press.
Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., & Craw-
ford, A. (2002). Internet paradox revisited. Journal of Social Issues, 58,
49 –74.
Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., &
Scherlis, W. (1998). Internet paradox: A social technology that reduces
social involvement and psychological well-being? American Psycholo-
gist, 53, 1017–1031.
Lizotte, A. J., Chard-Wierschem, D. J., Loeber, R., & Stern, S. B. (1992).
A shortened Child Behavior Checklist for delinquency studies. Journal
of Quantitative Criminology, 8, 233–245.
Loeber, R., Green, S. M., Lahey, B. B., & Stouthamer-Loeber, M. (1991).
Difference and similarities between children, mothers, and teachers as
informants on disruptive child behavior. Journal of Abnormal Child
Psychology, 19, 75–95.
Maccoby, E. E. (1998). The two sexes: Growing up apart, coming together.
Cambridge, MA: Harvard University Press.
Madden, M. (2006). Internet penetration and impact. Retrieved from
http://pewinternet.org
Mesch, G. S., & Talmud, I. (2007). Similarity and the quality of online and
offline social relationships among adolescents in Israel. Journal of
Research on Adolescence, 17, 455– 466.
Mitchell, K. J., Finkelhor, D., & Becker-Blease, K. A. (2007). Classifica-
tion of adults with problematic internet experiences: Linking internet
and conventional problems from a clinical perspective. CyberPsychol-
ogy & Behavior, 10, 381–392.
Modayil, M. V., Thompson, A. H., & Varnhagen, S. (2003). Internet users’
prior psychological and social difficulties. CyberPsychology & Behav-
ior, 6, 585–590.
Morahan-Martin, J., & Schumacher, P. (2003). Loneliness and social uses
of the internet. Computers in Human Behavior, 19, 659 – 671.
Papacharissi, Z., & Rubin, A. M. (2000). Predictors of internet use. Journal
of Broadcasting & Electronic Media, 44, 175–196.
Parker, J. G., & Asher, S. R. (1993). Friendship and friendship quality in
middle childhood: Links with peer group acceptance and feelings of
loneliness and social dissatisfaction. Developmental Psychology, 29,
611– 621.
Parks, M. R., & Roberts, L. D. (1998). Making MOOsic: The development
of personal relationships on line and a comparison to their off-line
counterparts. Journal of Social and Personal Relationships, 15, 517–
537.
Pedersen, S., Vitaro, F., Barker, E. D., & Borge, A. I. H. (2007). The timing
of middle-childhood peer rejection and friendship: Linking early behav-
ior to early-adolescent adjustment. Child Development, 78, 1037–1051.
Peter, J., Valkenburg, P. M., & Schouten, A. P. (2005). Developing a
model of adolescent friendship formation on the internet. CyberPsychol-
ogy & Behavior, 8(5), 423– 430.
Pew Internet and American Life Project. (2008). Demographics of internet
users. Retrieved from http://www.pewinternet.org/trends/User_
Demo_2.15.08.htm
Pew Internet and American Life Project. (2009). Generations online in 2009.
Retrieved from http://www.pewinternet.org/pdfs/PIP_Generations_
Memo.pdf
Sanders, C., Field, T., Diego, M., & Kaplan, M. (2000). The relationship of
internet use to depression and social isolation among adolescents. Ado-
lescence, 35, 237–242.
Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The
development of the person: The Minnesota study of risk and adaptation
from birth to adulthood.
Steer, R. A., Ball, R., Ranieri, W. F., & Beck, A. T. (1999). Dimensions of
the Beck Depression Inventory–II in clinically depressed outpatients.
Journal of Clinical Psychology, 55, 117–128.
Steiner, P. (1993, July 5). On the Internet, nobody knows you’re a dog
[cartoon]. The New Yorker, 69, 61.
Stocker, C. M., & Richmond, M. K. (2007). Longitudinal associations
between hostility in adolescents’ family relationships and friendships
and hostility in romantic relationships. Journal of Family Psychology,
21, 490 – 497.
Subrahmanyam, K., Greenfield, P., Kraut, R., & Gross, E. F. (2001). The
impact of computer use on children’s and adolescents’ development.
Applied Developmental Psychology, 22, 7–30.
TechCrunch. (2008). Facebook no longer the second largest social net-
work. Retrieved from http://www.techcrunch.com/2008/06/12/
facebook-no-longer-the-second-largest-social-network/
Timbremont, B., & Braet, C. (2004). Assessing depression in youth:
Relation between the Children’s Depression Inventory and a structured
interview. Journal of Clinical Child and Adolescent Psychology, 33,
149 –157.
Twenge, J. M., & Nolen-Hoeksema, S. (2002). Age, gender, race, socio-
economic status, and birth cohort differences on the Children’s Depres-
sion Inventory: A meta-analysis. Journal of Abnormal Psychology, 111,
578 –588.
Tyler, T. R. (2002). Is the internet changing social life? It seems the more
things change, the more they stay the same. Journal of Social Issues, 58,
195–205.
Valkenburg, P. M., & Peter, J. (2007). Preadolescents’ and adolescents’
online communication and their closeness to friends. Developmental
Psychology, 43, 267–277.
Valkenburg, P. M., Peter, J., & Schouten, A. P. (2006). Friend networking
sites and their relationship to adolescents’ well-being and social self-
esteem. CyberPsychology & Behavior, 9, 584 –590.
Walther, J. B., Van Der Heide, B., Kim, S. Y., Westerman, D., & Tong,
S. T. (2008). The role of friends’ appearance and behavior on evalua-
tions of individuals on Facebook: Are we known by the company we
keep? Human Communication Research, 34, 28 – 49.
Willoughby, T. (2008). A short-term longitudinal study of internet and
computer game use by adolescent boys and girls: Prevalence, frequency
of use, and psychosocial predictors. Developmental Psychology, 44,
195–204.
Wolak, J., Mitchell, K., & Finkelhor, D. (2003). Escaping or connecting?
Characteristics of youth who form close online relationships. Journal of
Adolescence, 26, 105–119.
Ybarra, M. L., Alexander, C., & Mitchell, K. J. (2005). Depressive symp-
tomatology, youth internet use, and online interactions: A national
survey. Journal of Adolescent Health, 36, 9 –18.
Received May 1, 2008
Revision received June 26, 2009
Accepted July 21, 2009 䡲